How to Sign In
Criminal Justice Blog
Showing results for 
Search instead for 
Did you mean: 
Emotion Detection: Fact or Fiction?
1 Comment

Emotion detection: the idea that a computer can read your facial expressions and body language to figure out how you're feeling. It's a compelling idea for software developers in a huge range of markets. So much so that some experts estimate the market for emotion detection software will be as much as $3.8 billion by 2025. That's rather tidy economic growth, especially for a field that's not exactly supported by a wealth of success.


What is Emotion Detection? 


Emotion detection purports to use technology to read people's emotional state based on objective physical input. By reading information like facial expressions, body language, eye tracking, heart rate, and respiratory rate, emotion detection software attempts to make a conclusion about an individual's actual emotions. The vast majority of emotion detection software uses facial expressions to draw conclusions about the emotions of the person in the eye of the camera. 


Does Emotion Detection Actually Work? 


Emotion detection can't work reliably, according to a group of researchers who just published their findings after reviewing thousands of psychological studies spanning several decades of emotion/physical expression research.

"Facial configurations...are not 'fingerprints' or diagnostic displays that reliably and specifically signal particular emotional states," the paper conclude:


"It is not possible to confidently infer happiness from a smile, anger from a scowl, or sadness from a frown, as much of current technology tries to do when applying what are mistakenly believed to be the scientific facts." 


Some facial expressions--like a smile--are fairly universal across all cultures. Other expressions are so ambiguous they could mean a wide variety of things: a frown could mean you're sad, frustrated, angry, or simply that you're concentrating. That's true for a huge range of individuals across the spectrum of cultural and circumstantial contexts. Even the meaning of a smile can be less than clear: it can be a grimace of pain or a sarcastic laugh. According to researchers, the level of variation across people's facial expressions and body language makes those features alone a notably unreliable method of predicting the emotions of a person on the other side of a camera lens. 


Emotion detection isn't a brand-new field. It operates on much the same premise as a polygraph (lie detector) test, by assuming that certain objective characteristics always hold true for the people in question. Unfortunately (or perhaps fortunately), it's not that simple. Physical movements, characteristics, and tendencies don't always correlate with a specific emotion or intention, and relying on them too much can be disastrous. 


This research flies in the face of a multi-million dollar industry churning out sparkling new emotion detection software for a massive array of software applications from video games to the medical field. What does the future of emotion detection hold? It's impossible to know for sure, but for now, the best emotion detector out there is still the imperfect, inconsistent one operating moment-by-moment in your own brain.



  1. Experts Say 'Emotion Recognition' Lacks Scientific Foundation
  2. Can you tell how people are feeling from their expressions? Quiz
  3. A Short Test On Recognizing Facial Expressions
  4. Introduction to Emotion Recognition

Related Article

1 Comment
Valued Contributor

This is a very interesting article. I recognize that while a computer may be unable to detect nuances of facial expressions to determine authentic moods, however, I disagree with the assertion that "Emotional detection can't work reliably." First, if there is a way to quantify "reliably" that would be a great debate in and of itself. Is 51% considered reliable? If I knew that I could roll the dice and I would win 51% of the time, I would go to Las Vegas and retire comfortably! But if reliable is considered 100%, then no; Emotional detection can't work reliably...then again very few things have a 100% reliability.

The paper 'concludes' that it's not possible to confidently infer happiness from a smile, et .al, but Dr. Paul Ekman "...discovered that several facial expressions of emotion, such as fear, anger, sadness, joy, and surprise were universal, and that people could easily read these expressions in people from different cultures." He shares in his training that it's not difficult to discern the emotional state of an infant or even the non-verbal through facial expressions. He used the term “micro-facial expressions” to explain that sometimes the authentic expression is quickly reflected on a person’s face before it is just as quickly replaced with something else. I would suggest that a simple test of his theory is to watch a news program or interview of a person with the sound turned off and focus on their facial expressions. It is surprising what you will see reflected just in the face!

Lastly, I believe that if there were an application that could incorporate an analysis of all the autonomic physical expressions, facial expressions, voice analysis, respirations, blood pressure, body temperature and body language against a baseline for the individual, then a more accurate assessment could be achieved. Of course the challenge would be obtaining a baseline for comparison.

Nonetheless, I find the idea of emotional detection to be a promising field and opportunity for law enforcement and other investigatory protective entities, TSA screeners, background investigators, Title IX Investigators etc. If such a system could be made mobile, you would have a “Truth Detection in a Box”!

How exciting is that!

Paul Ekman Biography